Abstract
OBJECTIVE: To investigate the survival outcomes and risk factors for mortality in cirrhotic patients with probable spontaneous bacterial peritonitis (SBP). METHODS: We retrospectively analyzed the clinical data of 323 cirrhotic patients with ascites admitted from June 2021 to May 2022, including 115 patients with SBP [ascites polymorphonuclear leucocyte (PMN) count ≥250/mm(3)], 52 patients with bacterascites (PMN count < 250/mm(3) with positive microbiological finding in ascites), 67 patients with probable SBP (PMN count < 250/mm(3) with negative microbiological finding in ascites but clinical symptoms of SBP) and 89 patients without infection (PMN count < 250/mm(3) with negative microbiological finding without clinical symptoms of SBP). The clinical characteristics, laboratory data and 90-day mortality of the patients were compared among the 4 groups. Cox proportional hazard model and propensity score matching (PSM) in a 1∶1 ratio were used to analyze the risk factors for mortality in patients with probable SBP. RESULTS: The patients with probable SBP had a 90-day mortality rate of 43.28%, similar to those of patients with SBP (46.95%, P=0.121) and bacterascites (48.07%, P=0.805) but significantly higher than that of non-infected patients (11.23%, P < 0.001). In the 46 pairs of patients matched using PSM, the 90-day mortality rates were higher in probable SBP group than in non-infected group both before (43.28% vs 11.23%, P < 0.001) and after PSM (34.78% vs 15.21%, P=0.038). Cox regression analysis indicated that probable SBP was an independent predictor of 90-day mortality in cirrhotic patients with ascites (HR=1.539, 95% CI: 1.048-2.261, P=0.028). A Model for End-Stage Liver Disease (MELD) score > 15 (HR=1.943, 95% CI: 1.118-3.377, P=0.018) and procalcitonin level > 0.48 ng/mL (HR=1.989, 95% CI: 1.111-3.560, P=0.021) at diagnostic paracentesis were both independent risk factors for 90-day mortality in patients with probable SBP. CONCLUSION: Cirrhotic patients with probable SBP have poor survival outcomes, and their management should be further optimized based on their MELD score and procalcitonin level.